AZ 28(1)-text_Layout112/8/1410:44AMPage101 Address forcorrespondence: Cullowhee, NC28723,USA. Western CarolinaUniversity, Department ofPsychology, E-mail: [email protected] ANTHROZOÖS Harold Herzog, Uncorrected OUETA SU B ERNSAALBEPHOTOCOPYING REPRINTSAVAILABLE PP. 0–001– VOLUME TBA,ISSUETBA David McCord is oneofthemostwidelyused measures ofgeneralattitudes toward behaviors(e.g.,meateating,involvement inanimalprotection). about animalsentience,and therelationship between attitudesand , personalitydifferences, socialclassandpoliticalideology, beliefs the impactsoffactorssuchaseducation,earlyexperiences withpetsor ethics canbeaddressed via attitudescales.Theseinclude,forexample, Many questionsrelated to thepsychologicalunderpinningsofanimal ual differences inattitudestoward theethicsofuseotherspecies. with ,andmuchlessattentionhasbeengiventoassessing individ- of thescalestheylocated,however, assessed aspectsofrelationships feeding tofrequency ofexperiences withanimalcruelty. Themajority(62%) actions. Theseinstrumentsrangedfrom measures ofbeliefsaboutbird examination of140English-languagemeasures ofhuman – inter- andNetting(2012)recently undertook acomprehensivespecies. Wilson brief scales Keywords: of attitudesrelated toanimal welfare. offer forresearchers alternatives whoneedconvenientandshortmeasures ity. TheAAS-5 andtheAAS-10haveexcellentpsychometricproperties and p versions were highlycorrelated withtheoriginal20-itemAAS(rs and a10-itemversionofthewidelyusedAnimalAttitudeScale(AAS).Both analysis andusinganexistingdataset(n tists inbriefmeasures ofattitudes andpersonalitytraits.Basedonfactor ABSTRACT † * Harold Herzog Attitude Scale Brief Measures oftheAnimal University, USA Department ofPsychology, Western CarolinaUniversity, USA California SchoolofProfessionalPsychology, AlliantInternational ❖ < 0.001),andbothversionsdemonstratedacceptablereliability andvalid- The AnimalAttitudeScale(AAS) designed tomeasure aspectsofourrelationships with other life,researchers havedevelopedavarietyofinstruments Due totheincreasing awareness oftheimportance in Animal AttitudeScale,animalethics,attitudemeasurement, There isincreasing interest amongsocialandbehavioralscien- H ULSESBYLICENSEONLY PERMITTED THE PUBLISHERS DIRECTLY FROM * * , StephanieGrayson 1 (Herzog, BetchartandPittman 1991) = 400adults),weconstructeda5- Proof † and PRINTED INTHEUK © ISAZ2015 > 0.95,

ft-101 Anthrozoös DOI: 10.2752/089279315X14129350721894 AZ 28(1)-text_Layout 1 12/8/14 10:44 AM Page 102 Uncorrected Proof

Brief Measures of the Animal Attitude Scale

animal protection. The original version of the AAS consisted of two subscales. The “Ethics Subscale” included 20 items which assessed attitudes to the treatment and use of animals (e.g., “It is unethical to breed purebred dogs for pets when millions of dogs are killed in animal shelters each year”). The “Take Action Subscale” consisted of nine items that focused on the degree to which people would be willing to engage in specific behaviors to save an animal or to facilitate (e.g., “I would be unlikely to stop my car to help an injured dog”). Factor analyses of the original version of the AAS, however, revealed that all 29 items fell onto a single factor. Hence, in subsequent administrations, “take action” items were dropped, and the present version of the AAS consists, with minor modifications in wording, of the 20 ethics subscale items. Originally developed for a study of sex role orientation and attitudes toward animal welfare issues, the AAS has subsequently been used to investigate other aspects of human–animal interactions. For example, the AAS has been used to study attitudes of German children toward animals (Binngießer, Wilhelm and Randler 2013), the impact of witnessing killing of animals (Daley and Morton 2008), the relationship between empathy and attitudes toward the use of animals (Taylor and Signal 2005; Daly and Morton 2009), attitudes of Chinese university students (Davey 2006), the relationship between ethical orientation and beliefs about animals use (Galvin and Herzog 1992), personality differences and animal attitudes (Mathews and Herzog 1997), dietary choices and views of animals (Ruby 2012), differences between animal protectionists and community members in Australia ( Signal and Taylor 2006), and the relationship between disgust sensitivity and animal welfare attitudes (Herzog and Golden 2009). The AAS has excellent psychometric . Cronbach’s alphas typically range from 0.85 to 0.95. The scale’s concurrent and convergent validity have also been established. AAS scores are positively correlated with the “Profit” subscale of Taylor and Signal’s (2009) , Pest, and Profit Scale (r = 0.83, p < 0.01) and negatively correlated with Grayson’s (2012) Scale (r = –0.93, p < 0.001). Animal protectionists score higher on the AAS than general community members (Taylor and Signal 2006), and members of animal use organizations score lower than people not involved in animal issues (Herzog and Golden 2009). The AAS predicts food choices, with vegetarians having significantly higher scores than non- vegetarians (Herzog and Golden 2009; Grayson 2012; Ruby 2012). AAS scores are also correlated with the importance that ethical concerns play in the decision of vegetarians to forgo eating animals (Herzog and Golden 2009). While the AAS is a psychometrically sound instrument for measuring attitudes toward the use of nonhuman species, some researchers have found the 20-item version overly long, par- ticularly in studies in which subjects take multiple scales in a single session. For example, in order to reduce the length of an online survey of attitudes toward animals of social workers, Faver and Muñoz (2014) created a 10-item version of the AAS. A 10-item adaptation of the AAS was used by Pearson, Dorrian and Litchfield (2011) to assess the effects of an environ- mental program in a setting. Bastian et al. (2012) used a 10-item version to study the effects of framing on moral concern for animals. A 6-item version of the AAS was recently used to study the effects of interacting with animals on the development of positive

Anthrozoös attitudes and behavior of children (Mueller 2014). While researchers have found brief versions of the AAS useful, different researchers have cre- ated their own versions of the scale. For example, the version Faver and Muñoz (2014) used to ft 102 study attitudes of social workers consisted of the 10 even-numbered items of the full AAS. We AZ 28(1)-text_Layout 1 12/8/14 10:44 AM Page 103 Uncorrected Proof

Herzog et al.

felt there was need for standardized brief measures of attitudes toward the use of non-human species. Hence, our objective was to use factor analysis to systematically develop brief (10-item) and very brief (5-item) versions of the AAS having acceptable psychometric properties. The use of shortened versions of established psychological measures has become com- mon in the behavioral sciences. For example, traditional five-factor personality measures (the “Big Five”) contain between 40 and 240 items. Gosling, Rentfrow and Swann (2003), however, demonstrated that the five major domains of personality can be measured with reasonable ac- curacy with only 10 items. In their Ten-Item Personality Inventory (TIPI), each personality di- mension is assessed by two questions. While the TIPI has slightly diminished psychometric properties, with over 1,300 citations in the past 10 years, it is widely accepted by personality researchers. In addition to personality, validity of brief scales has been established for other measures of individual differences. These include, for example, measures of work styles (Del Libano et al. 2010), perceptions of illness (Broadbent et al. 2006), the psychological state of “flow,” (Jackson, Martin and Eklund 2008), and depression (Henry and Crawford 2005). Note that concerns over the methodological limitations of shortened scales have been raised (see Konstabel et al. 2012). But as Smith, McCarthy and Anderson (2000) note, abbreviated forms of psychological scales can be developed that have acceptable psychometric properties. Anastasi (1988) stated that, generally speaking, longer tests are more reliable and valid that shorter tests; however, when a test is shortened by removing the least effective items, the shorter test may actually be an improvement.

Methods Participants The participants were a convenience sample of 400 American adults (60% male, 40% fe- male). They were a recruited through a variety of internet sources including Psychological Re- search on the Net, Lab-United, and through advertisements placed on websites such as Craigslist and Facebook. The mean age of the participants was 42 years (SD = 15.9). Forty- eight percent of participants were from suburban areas, 43% were urban, and 9% were rural. The median income of the sample was $66,000 (US dollars). Seventy-eight percent of the participants were Caucasian. As a group, they were highly educated: 57% had completed a graduate degree or attended graduate school and 24% had completed undergraduate de- grees. In terms of political affiliation, 47% were Democrats, 16% were Independent, 11% were Republican, 3% were members of the Green Party, and 17% had no political affiliation. The par- ticipants had diverse religious views: 16% were agnostic, 14% were Jewish, 12% were athe- ists, 10% were Catholic, 10% were Protestant, and 16% indicated no religious affiliation. The participants identified their diets as follows: (70%), vegetarians (10%), vegan (8%), piscatarian (6%), “omnivore with vegetarian leanings” (4%), “omnivore with ethical considera- tions” (1%), other (1%). Procedures The participants completed an on-line survey which included the 20-item Animal Attitude Scale, the 33-item Speciesism Scale, the 5-item Satisfaction with Life Scale (Diener et al. 1985), and a series of demographic questions. The Specieism Scale was developed by Anthrozoös Grayson (2012) to assess beliefs in human superiority and privilege over all other animal species. Sample items include “ are the most important species on earth” and “I am in favor of rights for non-human animals.” High scores on the scale indicate strong beliefs in ft 103 AZ 28(1)-text_Layout 1 12/8/14 10:44 AM Page 104 Uncorrected Proof

Brief Measures of the Animal Attitude Scale

human superiority. The Satisfaction with Life Scale measures global life satisfaction. It was included to assess divergent validity with the Speciesism Scale and was expected to be uncorrelated with Speciesism Scale scores. Construction of the AAS-10 and the AAS-5: As noted above, the AAS was revised previously to establish a single-dimension measurement instrument, the focal concept being “animal use ethics.” Principle components analysis (PCA) was conducted to verify unidimensionality within the current data set and to yield component loadings for each item on this dimension. The data clearly met assumptions for PCA: the sample size of 400 is considered good to very good for correlational techniques in general, and the ratio of subjects (400) to variables (20) exceeds the most stringent recommendations for factor analysis: there were no missing cases nor was there evidence of outliers; skewness (–0.137) and kurtosis (–0.780) were well within acceptable lim- its; results of the Kaiser-Meyer-Olkin Measure of Sampling Adequacy (0.953) and Bartlett’s Test of Sphericity (p < 0.001) were also clearly acceptable. The results of the PCA are shown in Table 1. Though three components had eigenvalues greater than 1.0, the results indicated that component #1 explained most of the variance (al- most 50%). A clear break occurred between components #1 and #2, and the remaining com- ponents trailed off gradually without a visible break. Further, examination of the component loadings for each item showed that all 20 items loaded most highly on component #1. All 20 had loadings greater than the traditional cut-off of 0.4, and 19 of the 20 had loadings greater than 0.6. Examination of item loadings on component #2 showed only a single item greater than 0.4, and there were only two items with loadings greater than 0.4 on component #3. (Note that none of these items was ultimately selected for either of the short forms.) Thus, we were comfortable in proceeding based on the assumption of a unidimensional parent scale. Items were chosen for the short forms using a hybrid approach in which psychometric prop- erties (primarily to enhance reliability) were balanced with item content (primarily to enhance convergent validity). As can be seen in Table 1, the 20 items on the parent form were first ranked by component #1 loading, in descending order. The direction of scoring (forward or reversed) was noted, as was the item content. To maximize internal consistency we selected items first based on component loading, but we also chose to balance scoring direction as well as to avoid redundant content. While “animal use ethics” is clearly the core construct measured by the AAS, the contexts in which it is applied cover a wide range of issues (e.g., breeding animals for food or fur, pet-keeping, hunting for food or sport, medical research, product safety research, and keeping animals in ). We realized that focusing solely on psychometric properties could very easily result in a final item set with limited contextual variation. This approach, while achiev- ing maximum reliability, would diminish convergent validity and the usefulness of scales as a measure of general attitudes toward the use of other species. Table 1, then, exhibits the exact decision-making process involved in item selection. Considering just the AAS-10, we can see that the first three items were chosen based on component loadings. The 4th and 5th ranked items were skipped as redundant (medical re- search), the 6th ranked was selected, the 7th was rejected as redundant, the 8th was se- lected, the 9th was rejected as redundant, the 10th was selected, and the 11th was rejected because at this point five of the items selected had been reverse-scored, so the remainder Anthrozoös needed to be forward-scored. The 12th item was rejected based on a unique content consideration. At this point in the item selection process, with four items left to select, major

ft 104 contextual domains had been covered, with several remaining: rodeos/circuses, zoos, hunting, AZ 28(1)-text_Layout 1 12/8/14 10:44 AM Page 105 Uncorrected Proof

Herzog et al.

Table 1. Criteria evaluated in selecting items for the short scales. Component/Loading Content Category Loading AAS-20 Scoring AAS-10 AAS-5 Rank # Item # 1 (2) (3) Direction Item # Item # 1 2 0.843 –0.255 0.068 R 2 2 Medical research 2 8 0.809 –0.194 –0.164 R 3 3 Food 3 14 0.790 0.155 –0.058 R 7 Skins (breeding for) 4 16 0.788 –0.387 0.043 R Medical research 5 7 0.773 –0.231 –0.048 Medical research 6 9 0.758 0.047 0.193 R 4 Human moral dominance

7 6 0.743 –0.208 0.003 R Food 8 19 0.719 –0.011 –0.120 10 Cosmetic safety testing 9 13 0.713 –0.071 0.449 R Human moral dominance 10 15 0.707 –0.359 –0.104 R 8 Dissecting for education 11 5 0.705 –0.269 –0.209 R Hunting (for food) 12 20 0.690 0.304 –0.294 Rodeos and circuses 13 18 0.686 0.280 0.151 R Food 14 11 0.668 0.321 –0.064 6 5 Zoos 15 1 0.660 0.324 –0.207 1 1 Hunting (for sport) 16 12 0.643 –0.057 0.468 R Human moral dominance 17 4 0.639 0.223 –0.236 Skins (trapping for) 18 17 0.634 0.109 –0.324 9 Pets 19 10 0.603 0.247 0.321 5 4 Slaughter of whales and dolphins 20 3 0.458 0.431 0.229 Cockfighting

pets, whales/dolphins, and cockfighting. We decided that zoos, hunting, pet keeping, and whale/dolphin slaughter were more salient contexts than rodeos/circuses, so we skipped this item and continued the previous strategy otherwise. The AAS-5 items were chosen based on a very similar approach, prioritizing the component loading, balancing the scoring direction, and covering major context domains in which considerations are relevant. The final items for the AAS-10 and AAS-5 are listed in Appendix 1.

Results Psychometric Properties Reliability and Validity: All three versions of the AAS had acceptable reliability. Cronbach’s alpha was 0.94 for the full AAS, 0.90 for the AAS-10, and 0.82 for the AAS-5. Correlations between

both brief scales and the full AAS were high: AAS-10 (r = 0.98, p < 0.001), AAS-5 (r = 0.95, Anthrozoös p < 0.001)). The convergent validity as measured by correlations of the brief scales with the Speciesism Scale in the same administration was also high: AAS-10 (r = –0.92, p < 0.001),

AAS-5 (r = –0.89, p < 0.001). ft 105 AZ 28(1)-text_Layout 1 12/8/14 10:44 AM Page 106 Uncorrected Proof

Brief Measures of the Animal Attitude Scale

Dietary Differences Two hundred and eighty-one of the participants characterized their diet as “omnivore” and 117 fell into the other dietary groups (“vegetarian,” “vegan,” “omnivore with ethical consider- ations,” etc). Because of the small numbers of participants in some of the dietary categories, for the purposes of analysis, we grouped the participants into two categories: “omnivore” and “other.” Self-identified “” scored lower on both brief versions than individuals in the “other” category: AAS-10 Omnivore M = 29.3 (SD = 7.7), “Other” M = 38.9 (SD = 9.5)

(t(396) = 9.678, p < 0.001, d = 1.110); AAS-5 Omnivore M = 16.4 (SD = 4.3), “Other” M = 22.0

(SD = 3.4) (t(396) = 5.587, p < 0.001, d = 1.447) Sex Differences Nearly all administrations of the 20-item AAS have found that women are more concerned with animal protection than men (e.g., Herzog, Betchart and Pittman 1991; Signal and Taylor 2007; Daly and Morton 2008; Binngießer, Wilhelm and Randler 2013). This was true of both brief versions of the AAS: AAS-10 Male M =28.9 (SD = 8.3), Female M = 34.3 (SD = 7.9)

(t(397) = 6.572, p < 0.001, d = 0.667); AAS-5 Male M = 16.2 (SD = 4.6), Female M = 19.3

(SD = 4.5) (t(397) = 6.649, p < 0.001, d = 0.681).

Discussion The AAS-10 and the AAS-5 are psychometrically robust short measures of attitudes toward the use of animals. Both the 5- and 10-item versions correlate very highly with the 20 item AAS. And both versions have acceptable reliability (Cronbach’s alphas > 0.80), though the reliability of the 10-item scale is somewhat better than that of the AAS-5. Credé et al. (2012) argue that use of one or two items to measure individual personality do- mains can compromise the validity of brief scales due to the increased probability of random measurement error. And because of content deficiency, they may have lower criterion validity. However, given the extremely high correlations between both short scales and the 20-item AAS, this criticism does not apply to the AAS-5 and AAS-10. Further, Credé et al. found using even slightly more items raised the reliability and validity of brief personality scales to the acceptable range. These findings also apply to our brief versions of the AAS in that 5 and 10 items rather than 1 or 2 items are used to measure a constellation of attitudes that has a single dimension. To conclude, the AAS-5 and the AAS-10 are reliable and valid measures of general attitudes toward the human use of other species. Our general belief is that, other things being equal, longer scales are preferable to brief and very brief forms. Hence, if time is not a constraint, researchers should consider using the long form of the AAS. If, however, participants are pressed for time or if the measurement of animal attitudes is embedded in a matrix of other measures, these brief forms of the AAS offer psychometrically sound measures of attitudes toward the use of other species by humans.

Note 1. The full 20-item AAS can be downloaded at: http://paws.wcu.edu/herzog/AnimalAttScale.pdf.

Anthrozoös References Anastasi, A. 1988. Psychological Testing. 6th edn. New York: McMillan. Bastian, B., Costello, K., Loughnan, S. and Hodson, G. 2012. When closing the human–animal divide expands

ft 106 moral concern: The importance of framing. Social Psychological and Personality Science 3(4): 421–429. AZ 28(1)-text_Layout 1 12/8/14 10:44 AM Page 107 Uncorrected Proof

Herzog et al.

Binngießer, J., Wilhelm, C. and Randler, C. 2013. Attitudes toward animals among German children and adolescents. Anthrozoös 26(3): 325–339. Broadbent, E., Petrie, K. J., Main, J. and Weinman, J. 2006. The Brief Illness Perception Questionnaire. Journal of Psychosomatic Research 60(6): 631–637. Credé, M., Harms, P., Niehorster, S. and Gaye-Valentine, A. 2012. An evaluation of the consequences of using short measures of the Big Five personality traits. Journal of Personality and Social 102(4): 874– 888. Daly, B. and Morton, L. L. 2008. Empathic correlates of witnessing the inhumane killing of an animal: An investigation of single and multiple exposures. & Animals 16(3): 243–255. Daly, B. and Morton, L. L. 2009. Empathic differences in adults as a function of childhood and adult pet ownership and pet type. Anthrozoös 22(4): 371–382. Davey, G. 2006. Chinese university students’ attitudes toward the ethical treatment and welfare of animals. Journal of Applied 9(4): 289–297. del Líbano, M., Llorens, S., Salanova, M. and Schaufeli, W. 2010. Validity of a brief workaholism scale. Psicothema 22(1): 143–150. Diener, E. D., Emmons, R. A., Larsen, R. J. and Griffin, S. 1985. The Satisfaction with Life Scale. Journal of Personality Assessment 49(1): 71–75. Faver, C. and Muñoz, J. D. 2014. Orientations to nonhuman animal welfare: A view from the border. Society & Animals 22(4): 372–389. Galvin, S. L. and Herzog, H. A. 1992. Ethical ideology, activism, and attitudes toward the treatment of animals. Ethics & Behavior 2(3): 141–149. Gosling, S. D., Rentfrow, P. J. and Swann Jr, W. B. 2003. A very brief measure of the Big-Five personality domains. Journal of Research in Personality 37(6): 504–528. Grayson, S. 2012. Measuring speciesism: Scale development and validation. Paper presented at the International Society for Anthrozoology conference, Cambridge, UK, July 11 to 13, 2012. Henry, J. D. and Crawford, J. R. 2005. The short-form version of the Depression Anxiety Stress Scales (DASS- 21): Construct validity and normative data in a large non-clinical sample. British Journal of Clinical Psychology 44(2): 227–239. Herzog, H. A. Betchart, N. S. and Pittman, R. B. 1991. Gender, sex role orientation, and attitudes toward animals. Anthrozoös 4(3): 184–191. Herzog, H. A. and Golden, L. L. 2009. Moral emotions and social activism: The case of animal rights. Journal of Social Issues 65(3): 485–498. Jackson, S. A., Martin, A. J. and Eklund, R. C. 2008. Long and short measures of flow: The construct validity of the FSS-2, DFS-2, and new brief counterparts. Journal of Sport & Exercise Psychology 30(5): 561–587. Konstabel, K., Lönnqvist, J. E., Walkowitz, G., Konstabel, K. and Verkasalo, M. 2012. The “Short Five”(S5): Measuring personality traits using comprehensive single items. European Journal of Personality 26(1): 13–29. Mathews, S. and Herzog, H. A. 1997. Personality and attitudes toward the treatment of animals. Society & Animals 5(2): 169–175. Mueller, M. K. 2014. Is human–animal interaction (HAI) linked to positive youth development? Initial answers. Applied Developmental Science 18(1): 5–16. Pearson, E., Dorrian, J. and Litchfield, C. 2011. Harnessing visual media in : Increasing knowledge of orangutan conservation issues and facilitating sustainable behaviour through video presentations. Environmental Education Research 17(6): 751–767. Ruby, M. B. 2012. . A blossoming field of study. Appetite 58(1): 141–150. Signal, T. D. and Taylor, N. 2006. Attitudes to animals in the animal protection community compared to a normative community sample. Society & Animals 14(3): 265–274. Signal, T. D. and Taylor, N. 2007. Attitude to animals and empathy: Comparing animal protection and general community samples. Anthrozoös 20(2): 125–130. Smith, G. T., McCarthy, D. M. and Anderson, K. G. 2000. On the sins of short-form development. Psychological Assessment 12(1): 102–111.

Taylor, N. and Signal, T. D. 2005. Empathy and attitudes to animals. Anthrozoös 18(1): 18–27. Anthrozoös Taylor, N. and Signal, T. D. 2009. Pet, pest, profit: Isolating differences in attitudes towards the treatment of animals. Anthrozoös 22(2): 129–135. Wilson, C. C. and Netting, F. E. 2012. The status of instrument development in the human–animal interaction ft 107 field. Anthrozoös 25(Suppl.): s11–s55. AZ 28(1)-text_Layout 1 12/8/14 10:44 AM Page 108 Uncorrected Proof

Brief Measures of the Animal Attitude Scale

Appendix 1. Animal Attitude Scale: 5- and 10-item versions. The AAS-10 consists of all the items below. The AAS-5 consists of the items in bold print. Higher scores indicate more concern for animal welfare. The numbers of points assigned to the response items are in parentheses. Starred items (**) are reverse scored. Instructions: Listed below are a series of statements regarding the use of animals. Circle the letters that indicate the extent to which you agree or disagree with the statement: SA = Strongly Agree (5), A = Agree (4), U = Undecided (3), D = Disagree (2), SD = Strongly Disagree (1) 1. It is morally wrong to hunt wild animals just for sport. SA A U D SD 2. I do not think that there is anything wrong with using animals in medical research.** SA A U D SD 3. I think it is perfectly acceptable for cattle and hogs to be raised for human consumption.** SA A U D SD 4. Basically, humans have the right to use animals as we see fit.** SA A U D SD 5. The slaughter of whales and dolphins should be immediately stopped even if it means some people will be put out of work. SA A U D SD 6. I sometimes get upset when I see wild animals in cages at zoos. SA A U D SD 7. Breeding animals for their skins is a legitimate use of animals.** SA A U D SD 8. Some aspects of biology can only be learned through dissecting preserved animals such as cats.** SA A U D SD 9. It is unethical to breed purebred dogs for pets when millions of dogs are killed in animal shel- ters each year. SA A U D SD 10.The use of animals such as rabbits for testing the safety of cosmetics and prod- ucts is unnecessary and should be stopped. Anthrozoös SA A U D SD ft 108